Issue Station
Sign In
Tools hereToolbox

transformers - 💡(How to fix) Fix Why max_shard_size in PreTrainedModel.save_pretrained() is 50GB? [1 comments, 2 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
huggingface/transformers#44484•Fetched 2026-04-08 00:28:11
View on GitHub
Comments
1
Participants
2
Timeline
3
Reactions
0
Author
SilencezongSilencezong
Participants
Rocketknight1Rocketknight1SilencezongSilencezong
Timeline (top)
closed ×1commented ×1labeled ×1
RAW_BUFFERClick to expand / collapse

System Info

In old version like 4.57.1, max_shard_size in PreTrainedModel.save_pretrained() is 5GB, but in new version, max_shard_size is '50GB'. Is it normal?

Who can help?

No response

Information

  • The official example scripts
  • My own modified scripts

Tasks

  • An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
  • My own task or dataset (give details below)

Reproduction

New version: def save_pretrained( self, save_directory: str | os.PathLike, is_main_process: bool = True, state_dict: dict | None = None, push_to_hub: bool = False, max_shard_size: int | str = "50GB", variant: str | None = None, token: str | bool | None = None, save_peft_format: bool = True, save_original_format: bool = True, **kwargs, ):

Old version: def save_pretrained( self, save_directory: Union[str, os.PathLike], is_main_process: bool = True, state_dict: Optional[dict] = None, save_function: Callable = torch.save, push_to_hub: bool = False, max_shard_size: Union[int, str] = "5GB", safe_serialization: bool = True, variant: Optional[str] = None, token: Optional[Union[str, bool]] = None, save_peft_format: bool = True, **kwargs, ):

Expected behavior

Fix this bug?

extent analysis

Fix Plan

Update max_shard_size Type

The root cause of the issue is the change in the type of max_shard_size from int to str in the new version of the save_pretrained method.

Step-by-Step Solution

  1. Update the type of max_shard_size in the new version of the save_pretrained method to int to match the old version:

def save_pretrained( ... max_shard_size: int | str = 50 * 1024 * 1024 * 1024, # 50GB in bytes ... ):

2. **Remove quotes from string values** in the `max_shard_size` parameter when calling the `save_pretrained` method:
   ```python
save_pretrained(save_directory, max_shard_size=50 * 1024 * 1024 * 1024)  # 50GB in bytes
  1. Update any hardcoded string values for max_shard_size to use the correct type:

max_shard_size = 50 * 1024 * 1024 * 1024 # 50GB in bytes save_pretrained(save_directory, max_shard_size=max_shard_size)


#### Verification

To verify that the fix worked, check that the `max_shard_size` parameter is correctly validated and used in the `save_pretrained` method. You can add print statements or use a debugger to inspect the value of `max_shard_size` within the method.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

FAQ

Expected behavior

Fix this bug?

#api#ssr#installation#tensor shape#autograd error#LLM response#prompt template#agent execution#callback error#memory management

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

transformers - 💡(How to fix) Fix Why max_shard_size in PreTrainedModel.save_pretrained() is 50GB? [1 comments, 2 participants]